Speaker: Warrick Ball, Institut für Astrophysik Göttingen

Abstract:
Oscillation frequencies of stellar models are known to show systematic differences with observations, owing to poor modelling of the near-surface layers.  These deviations prevent the exploitation of accurate space-based results from Kepler and CoRoT.  In this talk, I will present results from investigations of three methods for correcting the frequencies.  I first briefly discuss our attempts to fit stellar models to three-dimensional magnetohydrodynamic simulations of the near-surface layers.  I then examine a recently-proposed Bayesian method and show that it is potentially heavily biased.  Finally, I present a new, theoretically- motivated parametrization of the surface effect that fits the known differences between observed and modelled solar frequencies better than the current commonly-used power law.  I show early (and successful) tests for bias, and preliminary results for fits to a real star, HD 52265.